Fetching the paper…

Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs · Around